Adoption of artificial intelligence in smart animal husbandry in Qinghai pastoral areas: the driving mechanism of technological traits and user beliefs

The rapid development of agricultural artificial intelligence (AI) supports the transformation of traditional to smart animal husbandry, yet its adoption in pastoral areas is constrained by complex geographical environments and uneven technological cognition. This study explores pastoralists’ AI adoption intention using a framework integrating technological innovation characteristics (mobility, interactivity, autonomy), technological commitment (technological interest, competence belief, control belief), and the Technology Acceptance Model (TAM), with animal husbandry type and breeding scale as moderators. Partial Least Squares Structural Equation Modeling (PLS-SEM) analyzed 321 valid questionnaires. Results show: (1) Technological innovation characteristics: Mobility ( β = 0.375, p < 0.005), interactivity ( β = 0.193, p < 0.05), and autonomy ( β = 0.122, p < 0.05) positively influence perceived usefulness; all three also enhance perceived ease of use (mobility: β = 0.164, p < 0.005; interactivity: β = 0.273, p < 0.005; autonomy: β = 0.210, p < 0.005). (2) Technological commitment: Technological interest promotes perceived usefulness ( β = 0.230, p < 0.005) and ease of use ( β = 0.169, p < 0.05), while competence and control beliefs show no significant effects. (3) Perceived usefulness ( β = 0.320, p < 0.001) and ease of use ( β = 0.245, p < 0.001) directly drive adoption intention. Moderating effects: Animal husbandry type positively moderates autonomy-perceived ease of use ( β = 0.086, p < 0.05) and technological interest-perceived usefulness ( β = 0.099, p < 0.05), but negatively moderates interactivity-perceived usefulness ( β = −0.220, p < 0.005). Breeding scale positively moderates technological interest’s effects on perceived usefulness ( β = 0.096, p < 0.05) and ease of use ( β = 0.109, p < 0.05). This study reveals differentiated AI adoption paths in Qinghai pastoral areas, providing a theoretical basis for targeted strategies. Recommendations include optimizing AI design by animal husbandry type, delivering hierarchical training by breeding scale, and strengthening technological competence and control belief cultivation to improve AI adaptability and sustainability.